Abstract
BACKGROUND: Promising cancer treatments, such as DDR inhibitors, are often challenged by the heterogeneity of responses in clinical trials. The present work aimed to build a computational framework to address those challenges. METHODS: A semi-mechanistic pharmacokinetic-pharmacodynamic model of tumour growth inhibition was developed to investigate the efficacy of PARP and ATR inhibitors as monotherapies, and in combination. Key features of the DNA damage response were incorporated into the model to allow the emergence of synthetic lethality, including redundant DNA repair pathways that may be impaired due to genetic mutations, and due to PARP and ATR inhibition. Model parameters were calibrated using preclinical in vivo data for PARP inhibitors rucaparib and talazoparib and the ATR inhibitor gartisertib. RESULTS: The model successfully captured the monotherapy efficacies of rucaparib and talazoparib, as well as the combination efficacy with gartisertib. The model was evaluated against multiple tumour xenografts with diverse genetic backgrounds and was able to capture the observed heterogeneity of response profiles. CONCLUSIONS: By enabling simulation of in vivo tumour growth inhibition with PARP and ATR inhibitors for specific tumour types, the model provides a rational approach to support the optimisation of dosing regimens to stratified populations.